304
Q. Chen et al.
As such, with the high differences in overlap radio were so great between SVP and
RSVP (as opposed to the inter-reader difference), the RSVP method was concluded
to be superior when compared to conventional SVP.
In conclusion, we developed a new method to improve visualization of drusen on
an RPE-based projection of 3D OCT retinal images. This method uses automated
RPE segmentation, drusen segmentation, and image post-processing to enhance the
conspicuity of drusen on the projection image and to minimize the amount of extraneous retinal tissue contributing to the projection. Using quantitative evaluation analysis, by comparing RSVP and conventional SVP images against a gold standard, the
RSVP method was evidently more effective for drusen visualization, which may be
useful to ophthalmologists in directly and rapidly assessing the macula of patients
who have non-exudative age-related macular degeneration.
11.3 Geographic Atrophy Segmentation and Visualization
Several semi-automatic and automatic GA segmentation methods [45, 46] have been
proposed for FAF images. A region-growing method was proposed by Deckert [47],
where separate GA regions needed to be manually seeded to be included in the segmentation. Lee [48] adopted a level set model, and proposed a hybrid approach by
identifying hypo-fluorescence GA regions from other interfering vessel structures in
the FAF images [49] and an interactive segmentation approach by using the watershed
transform algorithm [50]. Sayegh [51] evaluated SD-OCT for grading GA compared
with FAF images, and concluded that SD-OCT is an appropriate imaging modality
for evaluating the extent of GA lesions. Chiu [52] used graph theory and dynamic
programming to segment retina layers in eyes with GA and drusen. Schütze [53]
suggests that the current available automated segmentation methods are limited in
their ability to accurately assess retinal layer thickness and are thus not accurate in
detecting GA. At present, if quantitative assessment of GA in SD-OCT images is
desired, it needs to be performed by an expert who manually circumscribes the GA
lesions in the B-scan images (the primary output from an SD-OCT device, comprising 2D contiguous slices through a volumetric cube of the retina), and subsequently
projecting the segmentations onto an en face image to show the extent of GA across
the retinal surface—a similar view to that seen in FAF images. Each SD-OCT volumetric image dataset generally contains 128 or 200 B-scan images (for CirrusOCT
(Carl Zeiss Meditec, Inc., Dublin, CA)). Since this manual circumscription of GA
lesions in the B-scans is very time-consuming, it is not routinely performed in clinical
practice. Other methods are also proposed recently by researchers [54–56].
This section presents two novel GA segmentation and two visualization algorithms
for SD-OCT images, namely; (a) Semi-automatic geographic atrophy segmentation
for SD-OCT images [57], (b) Automated GA segmentation for SD-OCT images
using CVLSF model [58], (c) Restricted summed-area projection for geographic
atrophy visualization in SD-OCT images [59] and (d) A false color fusion strategy
for drusen and GA visualization in OCT images [60].
Q. Chen et al.
As such, with the high differences in overlap radio were so great between SVP and
RSVP (as opposed to the inter-reader difference), the RSVP method was concluded
to be superior when compared to conventional SVP.
In conclusion, we developed a new method to improve visualization of drusen on
an RPE-based projection of 3D OCT retinal images. This method uses automated
RPE segmentation, drusen segmentation, and image post-processing to enhance the
conspicuity of drusen on the projection image and to minimize the amount of extraneous retinal tissue contributing to the projection. Using quantitative evaluation analysis, by comparing RSVP and conventional SVP images against a gold standard, the
RSVP method was evidently more effective for drusen visualization, which may be
useful to ophthalmologists in directly and rapidly assessing the macula of patients
who have non-exudative age-related macular degeneration.
11.3 Geographic Atrophy Segmentation and Visualization
Several semi-automatic and automatic GA segmentation methods [45, 46] have been
proposed for FAF images. A region-growing method was proposed by Deckert [47],
where separate GA regions needed to be manually seeded to be included in the segmentation. Lee [48] adopted a level set model, and proposed a hybrid approach by
identifying hypo-fluorescence GA regions from other interfering vessel structures in
the FAF images [49] and an interactive segmentation approach by using the watershed
transform algorithm [50]. Sayegh [51] evaluated SD-OCT for grading GA compared
with FAF images, and concluded that SD-OCT is an appropriate imaging modality
for evaluating the extent of GA lesions. Chiu [52] used graph theory and dynamic
programming to segment retina layers in eyes with GA and drusen. Schütze [53]
suggests that the current available automated segmentation methods are limited in
their ability to accurately assess retinal layer thickness and are thus not accurate in
detecting GA. At present, if quantitative assessment of GA in SD-OCT images is
desired, it needs to be performed by an expert who manually circumscribes the GA
lesions in the B-scan images (the primary output from an SD-OCT device, comprising 2D contiguous slices through a volumetric cube of the retina), and subsequently
projecting the segmentations onto an en face image to show the extent of GA across
the retinal surface—a similar view to that seen in FAF images. Each SD-OCT volumetric image dataset generally contains 128 or 200 B-scan images (for CirrusOCT
(Carl Zeiss Meditec, Inc., Dublin, CA)). Since this manual circumscription of GA
lesions in the B-scans is very time-consuming, it is not routinely performed in clinical
practice. Other methods are also proposed recently by researchers [54–56].
This section presents two novel GA segmentation and two visualization algorithms
for SD-OCT images, namely; (a) Semi-automatic geographic atrophy segmentation
for SD-OCT images [57], (b) Automated GA segmentation for SD-OCT images
using CVLSF model [58], (c) Restricted summed-area projection for geographic
atrophy visualization in SD-OCT images [59] and (d) A false color fusion strategy
for drusen and GA visualization in OCT images [60].